- Email: [email protected]
Performative design environment for kinetic photovoltaic architecture
Automation in Construction 93 (2018) 339–347
Contents lists available at ScienceDirect
Automation in Construction journal homepage: www.elsevier.com/locate/autcon
Performative design environment for kinetic photovoltaic architecture P. Jayathissa a b
a,*,1
, S. Caranovic
a,1
a
b
, J. Hofer , Z. Nagy , A. Schlueter
T
a
Architecture and Building Systems, Institute of Technology in Architecture, Department of Architecture, ETH Zurich, Switzerland Intelligent Environments Laboratory, Department of Civil, Architectural and Environmental Engineering, The University of Texas at Austin, USA
A R T I C LE I N FO
A B S T R A C T
Keywords: Kinetic architecture Multi-functional envelope Parametric design Adaptive shading Performative design
The design of complex architectural components such as kinetic architectural elements poses a challenge due to the multiple technological branches involved. This paper presents a performative design environment that combines the branches of structural and energy engineering, control, industrial design, and architecture. The methodology is applied in the context of the Adaptive Solar Facade, a kinetic photovoltaic shading system for the HiLo building in Duebendorf, Switzerland. The authors describe how the environment enables the user to design the form of the facade, get feedback on its structural strength, analyse the energetic performance of the interior space, conduct a daylighting analysis, render images, and produce fabrication plans for a rapid design process. With the parametric design environment, project meetings transform from information exchanges to design sessions where all stakeholders can collaboratively influence the design and see immediate results. What would normally take a month, was condensed to just a few hours. Ultimately, this paper extends the field of performative design by presenting a practical example where a system as complex as a kinetic photovoltaic envelope can be designed, prototyped, and fabricated by a small team of four designers.
1. Introduction The interest for applying kinetic envelopes in architecture has increased in recent years. On the one hand, there is the availability of digital design processes, which enables kinetic concepts to come to life. On the other hand, there is the desire to have active building elements to regulate energy flows and to create a more comfortable built environment [1]. When analysing the various energy flows, it is solar radiation that is of particular interest. As seen in Fig. 1, the mediation of solar radiation has the potential to reduce heating and cooling demands, while simultaneously distributing daylight according to the occupants' desires [2]. Furthermore, utilising thin film photovoltaic panels as kinetic elements enables the facade to also act as an electricity generator. Such technologies can, in some cases, offset the entire energy consumption of the rooms behind the envelope [3]. The application of kinetic architectural envelopes has so far been centred around iconic examples. These include the Al Bahar Towers in Dubai, the Arab World Institute in Paris, and the ThyssenKrupp Headquarters in Essen. Bringing these technologies to the mainstream, however, can be challenging.
One major challenge in the design of Kinetic Architectural Elements is the involvement of multiple technological branches. Among these branches one can count structural and energy engineering, control, industrial design, and architecture. Each of these branches has strong interdependencies to each other. For example, during the detail design process, a small variation in the control system, such as the range of movement, can affect the energetic performance and the architectural image. This may result in a redesign, that then needs to undergo structural evaluations and fit within the budget. These iteration cycles become a time intensive process that, in some cases, can take months to solve. In order to unite these branches efficiently, new design methods and environments have to be developed. The state of the art in interdisciplinary building design lies in building information modelling (BIM) [4,5]. The utilisation of BIM to perform fast energy and structural assessments can coordinate the design from its early stages. However, design using BIM is often based on high levels of standardisation, which makes it complicated when designing custom innovative components. The development of kinetic architectural envelopes requires a flexible infrastructure, which allows for fast design, prototyping and production, while maintaining the ability to be customised for each system individually.
Abbreviations: ASF, Adaptive Solar Facade; BIM, Building information modelling; CAD, Computer aided design; PDE, Performative design environment; PV, Photovoltaic; FEA, Finite element analysis * Corresponding author. E-mail addresses: [email protected] (P. Jayathissa), [email protected] (S. Caranovic), [email protected] (J. Hofer), [email protected] (Z. Nagy), [email protected] (A. Schlueter). 1 Equally contributing first authors. https://doi.org/10.1016/j.autcon.2018.05.013 Received 13 October 2017; Received in revised form 16 April 2018; Accepted 9 May 2018 0926-5805/ © 2018 Elsevier B.V. All rights reserved.
Automation in Construction 93 (2018) 339–347
P. Jayathissa et al.
shape the form of a facade, provide feedback on its structural strength, analyse the energetic performance and day-lighting conditions of the building space, render images, and produce fabrication drawings for a rapid iterative design process. This is accomplished within the Rhino/ Grasshopper environment with python as a scripting language. The methodology is applied in the context of the Adaptive Solar Facade (ASF), a kinetic photovoltaic facade constructed for the research and innovation unit known as the HiLo [13]. HiLo is a two bedroom apartment with a portfolio of energy saving technologies that create a net zero energy building [14]. Both bedrooms of the HiLo will be equipped with an ASF as seen in Fig. 2. The adaptive control of solar radiation into the bedrooms, coupled with on site electricity generation will contribute to this overall net zero energy strategy. The entire design process from conceptual design, to prototyping, and final fabrication will be presented. The remainder of the paper is organised as follows: the next section details the parametric design environment and the simulations that drive the final design. Section 3 describes the outcomes of this design process including the details of the kinetic facade, and in Section 4 the limitations of the environment are discussed and the paper is concluded.
importance max PV
ASF min
daylighting
views
Cooling
ASF Fig. 1. The facade acting as a mediator between the interior and exterior environment, while fulfilling various functions [2,15].
Performative design environments can overcome the limitations of BIM by taking computer aided design (CAD) in a reversed manner, where it is the simulations that drive the design [6]. Here, the concept of form making, is replaced with form finding. An example of performative design has been described by Turrin et al. where passive solar strategies were explored to improve the thermal comfort and daylight quality under a tessellated roof [7]. This can also stretch to structural performative design in tools such as RhinoVault where the final form is determined through iterative structural simulations [8]. Holzer on the other hand used performative design for direct structural feedback using first and second order structural simulations [9]. These tools can also be utilised on the component level, and have been previously analysed in a design studio for smart building envelopes [10]. However, integrating multiple tools in a single automated environment has still proven to be difficult due to some tools lacking parametrisation capabilities, low openness of the tools' interfaces, and low flexibility [11,12]. Furthermore, this lack of interoperability results in long iteration cycles, making it difficult to evaluate the necessary trade-offs. In this paper, the authors build on this existing knowledge to produce a state of the art performative design environment (PDE) that can
2. Methodology The parametric design environment, shown schematically in Fig. 3 details the multiple processes that were used in parallel to handle the different technological branches. From this, four key outputs are generated: the structural performance, energetic performance, manufacturing plans, and visual renders. Inputs to the PDE are defined by the parameters that have the greatest influence on the design. In this case, they are the overall frame dimension and profile, the photovoltaic panel dimension, spacing and layout, and the range of motion. These inputs are numerically fed into the design environment and generate instantaneous results of the structural performance, energetic performance, visual renders and manufacturing plans. By doing so, the multiple trade-offs between the technological branches, as explained in Section 1, can be simultaneously observed. The electricity generation, building energy demand, utilisation factor of yield strength,
Fig. 2. Render of the HiLo module to be constructed in Dubendorf, Switzerland. Two adaptive solar facades have been planned on the south and west facing facades. 340
Automation in Construction 93 (2018) 339–347
P. Jayathissa et al.
Fig. 3. The performative design environment is able to link three analysis methodologies, resulting in rapid design iterations. Here the panel spacing input parameter of the environment is varied. Once a design has been selected, production plans can be automatically generated leading to rapid physical implementation.
structural and energetic analysis in more detail.
dimensions, collision detection, aesthetics, and manufacturing costs are of particular interest. This ultimately allows for quick feedback loops with the major stakeholders (architects, structural engineering, energy engineers, and production team) involved in the project. The environment combines the geometric modelling software Rhinoceros 3D [16], its parametric plug-in Grasshopper [17], and python [18] as a scripting language. The relatively unspecialised nature of Rhino is complemented by a large number of specialised add-ons for Grasshopper. Furthermore, custom Grasshopper modules can be scripted using python. The following subsections will explain the
2.1. Energy evaluation The purpose of the ASF is to maximise electricity production on the photovoltaic (PV) panels and minimise the energy consumption of the building behind the facade. In order to best achieve this, the layout of the PV panels, and the electrical interconnection of PV panels must be carefully designed. The evaluation of the energetic performance of the ASF can be 341
Automation in Construction 93 (2018) 339–347
P. Jayathissa et al.
panels enables more control over the solar insolation. Tight configurations tend to perform better in hot climates, whereas sparse configurations perform better in colder climates. 4. Daylighting model: A linear daylighting model based on the total flux methodology is used to determine the luminosity in the room [22]. When the luminosity falls below a threshold value of 300lx, artificial lighting is turned on. Sparse configurations result in more daylight distribution, which reduces the need for artificial lighting in the morning and evening hours. 5. Optimisation: The simulation is conducted for all possible panel angle combinations for every hour of the year. By summing all the time steps, the annual energetic performance of the ASF can be evaluated.
Fig. 4. A simulation result showing the radiation on the solar panels and the window element on 16 June 2013 between 12:00 and 13:00 in Zurich.
This analysis finds the optimum balance between PV generation and daylight control to minimise heating, cooling and lighting demands where the overall objective is the minimisation of net energy. The source code for this methodology, with installation instructions can be downloaded from github [23,24]. During the design stage, this analysis is conducted for a typical day in summer, and in winter. Once a design has been selected, an annual study with hourly time steps is conducted to achieve a high resolution result.
found in [3], and will be briefly reviewed here for completion. This part of the PDE consists of five stages: 1. Solar radiation model: The radiation on the PV panels and window behind the ASF is calculated using the Grasshopper - LadyBug plugin. [19]. An example of the simulation result is shown in Fig. 4. A tighter layout of panels allows for more PV material per facade area, however it also results in more module self-shading, which reduces the overall electricity production. 2. PV electricity production: The radiation result on the PV panels is coupled to an electrical circuit simulation of monolithically interconnected, thin-film CIGS PV modules. This model takes into account the effects of module self-shading and temperature dependence [20]. 3. Building energy model: The radiation calculated on the window surface is fed to a 5R1C single zone resistance-capacitance building model based on the ISO 13790 standard [21]. This calculates the heating or cooling demand of the building. A tight layout of PV
2.2. Structural form finding and finite element analysis The panel spacing, as discussed in Section 2.1 can also influence the structural performance of the system, which in turn, influences the architectural image. It is therefore important to run the structural analysis in parallel to find the optimum solution. This subsection will first introduce the structural concept, and then detail how this concept was developed with the PDE. The proposed ASF is based on a vaulted cross-hatched network of stainless steel pipes as seen in Fig. 5. The junctions where the pipes
Fig. 5. The Adaptive Solar Facade existing in varying states. 342
Automation in Construction 93 (2018) 339–347
P. Jayathissa et al.
Fig. 6. An example output from the Karamba structural simulation. The black arrows detail the loading direction, and the colour details the utilisation factor in relation to the von Mises stress. The table on the right details the input parameters to the simulation.
Fig. 7. Effect of annual PV production with respect to module spacing per square meter of facade area [20].
cross serve as the mounting points for each dynamic photovoltaic module. All utility lines are routed within the pipe network. A steel frame supports the pipe network to create a stand-alone pre-fabricated component, which can be mounted directly to the building envelope. The vaulted shape further strengthens the structure against wind loads, thus allowing for thinner pipe diameters, which increases transparency. The form of the vaulted structure is determined through RhinoVault, a form finding plug-in for Grasshopper [8]. The method calculates the optimum shape of the pipe network based on the loading points and inputs to the design environment. Once the form is determined, a structural second order finite element analysis is conducted using the Karamba3d plug-in [25] to dimension the structural elements. Further manual adjustments to the mesh can be conducted to improve the architectural image. Each manual adjustment is directly computed by a second order finite element analysis (FEA), providing real time feedback about the stability of the adjusted structure. Fig. 6 shows an example of this output with a list of simulation parameters. A load of 420N was applied on each panel node, which is equivalent to a category one hurricane on the Saffir-Simpson scale. The results are visualised in the form of coloured meshes which detail the utilisation factor in relation to the von Mises stress. A scaled down 2.2 m × 1.5 m prototype was constructed to validate the structural model and the design of the connection details.
Fig. 8. Utilisation factor, represented as a percentage of the maximum yield strength. Failure occurs with utilisation factors greater than 100%. The utilisation factor decreases with increasing frame size cross section, and with greater vaulted depth.
2.3. Integration of classical design methods A purely parametric approach is very useful in investigating design possibilities, but once the concept needs to be put in practice, the parametric design environment needs to be complemented with 343
Automation in Construction 93 (2018) 339–347
P. Jayathissa et al.
Fig. 9. Prototype constructed for model validation. a, b details the experimental set up. c, d detail the snap-through buckling of the structure. Table 1 Comparison of the physical and simulated model for the scaled down prototype.
Max point loading Max deflection
Physical model
Simulated model
Deviation
250 N 14.5 mm
220 N 10.3 mm
−12% −29%
Fig. 11. Exploded view of the final ASF module.
physical prototypes, testing and classical design methods like simple 3D modelling and electronics design. As there is a trade-off between the flexibility of a parametrised environment and the time required for its development, it is important to wisely choose what to parametrise and what to design with classical methods. Parametrising the entire design process of complex systems such as an ASF can result in instabilities and lead to an increase in design time. For this reason, it is important to include certain constants amongst the parametric variables. Constants include the design of the electronics, connection details, and the actuator design [26]. Fig. 10. Constraints of panel motion relative to the structure.
3. Results This section details some of the outcomes of the aforementioned 344
Automation in Construction 93 (2018) 339–347
P. Jayathissa et al.
3.2. Optimum supporting frame dimension With a frame size of 2.5×4.02 m, a PV panel spacing of 510 mm and a structural depth of the vault of 150 mm, the shape of the rod-net vault, the profiles of the pipes, and the size of the supporting frame were calculated. With a wind load of 0.92 KN /m2, and a safety factor of 1.8, the finite element analysis concluded that the stainless steel pipe and the rectangular steel frame have a minimum dimension of 16×2 mm and 180× 120×6 mm respectively. The relatively large frame is the result of a snap-through buckling failure criterion and relatively weak support conditions at the corners. As the vault is loaded, the large lateral forces on the frame result in deflections, which in turn reduce the depth of the vaulted structure. The reduced depth of the vault further increases the lateral loading, thus leading to a global snap-through buckling failure. The frame must therefore be large enough to withstand the maximum load criterion with minimal deflection. This solution is characteristic to the conditions present on the HiLo building, where the ASF will be mounted with an offset of 80 cm from the building's support structure. In principle, the edge points of the pipe structure could be mounted directly to the support structure of the building, thus reducing the need for a frame. Fig. 8a depicts how the utilisation of the pipe elements decreases with increasing frame strength, up to the point where all edge points are supported individually. The 150 mm depth of the vaulted structure is sufficient to handle the wind loads with relatively thin pipes and minimal variation in the panel orientation. Ultimately, the choice of the structural depth is a trade-off between stability, self-shading, and aesthetics. A deeper structure offers more load bearing capacity requiring thinner pipes, however the self-shading of the panels increase. On the other hand, a shallow structure is less resilient to wind loads, leading to thicker pipes. The effect of the depth on the structural stability is depicted in Fig. 8b. The figure shows how the maximum utilisation of the pipes decreases with increasing structural depth, with a very steep decrease in the range of 12–15 cm. The construction details contribute significantly to the overall stability of the structure and to the match between the real behaviour and the FEA simulation. For this reason, the authors designed the pipe and frame connections to provide a completely stiff connection under the assumed wind loads. To form the vaulted shape, each pipe element is bent close to its edges and kept straight in between those points to improve their stability to localised buckling.
Fig. 12. Automatically generated drawings for pipe bending. Here four of the 92 pipe drawings are shown.
design environment in relation to the case study at the HiLo building. In particular the optimal PV panel layout, the structure, details of the module design, and fabrication will be evaluated. A typical iteration of the PDE would take 1 min. However a high resolution hourly analysis of the energetic performance with all solar angles can take up to 6 h. During a typical meeting, the simplified model would be run with approximately 20 iterations. Afterwards, a high resolution model is run for validation.
3.1. Optimum PV panel layout The design of the PV panel layout is an example where all four stakeholders had inputs to the design. From an energetic perspective, a dense layout would have a larger overall PV surface area for electricity generation. However if the panels are too close together, there would be high module self shading which would lower the overall performance of the panels, as seen in Fig. 7 [20]. Furthermore, there would be less natural lighting in the room, resulting in an increase of the building's overall energy consumption. From an architectural perspective, a sparse PV layout is preferred as it increases the transparency of the facade from the inside. This also lowers the overall cost, as less PV panels are required. Structurally, however, a sparse PV layout results in longer pipe elements between the junctions, which in some configurations, decreases the overall strength due to an increased buckling length. Ultimately, a PV panel spacing of 510 mm with a module size of 420 mm was chosen as a trade-off between the above-mentioned requirements.
3.3. Structural validation Due to the complexity of the joints and double curved form, it was
Fig. 13. Mounting of a prefabricated ASF to the test site. 345
Automation in Construction 93 (2018) 339–347
P. Jayathissa et al.
design session where all stakeholders can collaboratively influence the design and immediately see the necessary results. With traditional methods, five design iterations would normally take a month, whereas with the PDE, this can be condensed to a few hours. The design environment also develops with the project allowing for new parametric inputs, or new outputs to be created. The management of such a parametric design environment can be easily done with collaborative software management tools such as git, a distributed version control system. One disadvantage is the overhead required to manage a PDE. Like a BIM manager, a PDE manager is required and must have a careful overview of the software. As the software ultimately determines the final form of the design, any errors in the software can be detrimental to the final design. It was therefore necessary for all stakeholders to conduct a final independent analysis prior to the submission of the final design. One limiting factor in the design of the PDE is the computational time. The full annual energetic analysis, for example, may take 6 h to solve. Simplifications were therefore made to accelerate this process during the design stage, and the full complex evaluation was conducted afterwards to validate the simplified model. It is also important to determine what aspects of the design should be contained within the PDE and what should be designed with classical methods. Essentially, all aspects where there could be conflicts between the stakeholders were included in the PDE, whereas many of the design details, such as the mounting brackets to the building, can still be designed independently and imported into the PDE as a static object. Over time, some static objects, such as the cantilevered bracket of the PV module became parametrised. Ultimately, this paper presents a further step in the field of performative design by showing how a system as complex as a kinetic photovoltaic envelope, can be developed, prototyped and finally fabricated by a small team of four designers. The methodology can be utilised for any building component where multiple technological branches are required.
important to validate the structural analysis. A scaled down, 2.2 m× 1.5 m, prototype designed through the PDE was constructed for testing purposes. Weights were added to each of the seven junction nodes in 5 kg increments until the structure collapsed as seen in Fig. 9. Each junction node was also fitted with a deflection gauge. The results of this analysis are summarised in Table 1. The simulated model underestimates the overall strength by 12% which is sufficient to validate our model. The deflection results, however, have a larger diversion. This diversion can be attributed to a weak construction joint at one of the edge nodes. The average deflection of the structure is 10.5 mm which is close to the simulated model. Besides confirming the accuracy of the simulation, this test also confirmed the predicted snapthough buckling failure. Additionally, as seen in Fig. 9d, the junction nodes, and frame connections were capable of resisting local torsion and maintained their position. The failure was localised at the pipes as predicted in Section 3.2. 3.4. Module design The kinetic PV module was then parametrically designed to fit within the requirements of the optimised layout and structure. The PV panels must be able to fully open, and fully close without making contact with the rod-net structure or other PV modules as shown in Fig. 10. The PV panel is actuated using a soft pneumatic actuator [26]. The actuator is made from neoprene rubber and contains three air chambers. By pumping one or more of these chambers with compressed air, the actuator will deform, thus moving the panel with a 90° range in two degrees of freedom. The cantilevered bracket connects this actuator to the rod-net structure and holds it at a distance that prevents collision with other PV panels or the structure. A decentralised control box, located behind the junction element, contains three pneumatic valves and an electronic board addresses the valves over a common data bus. This controls the flow of air to the chambers inside the actuator. An exploded view of the module is shown in Fig. 11. 3.5. Fabrication
Acknowledgements Once a design has been configured using the PDE, manufacturing plans can be automatically generated. This enables the plans to be immediately sent to manufacturers, without a significant time investment. This is especially important for components such as the steel pipes where each pipe has a unique length and bend angle. As an example, Fig. 12 details four of the 92 pipe bending plans. Besides the plan, the PDE also generates a set of meta data that can be directly fed into a computer controlled pipe bending machine, thus simplifying the transition from design to production. An ASF for the HiLo building was fabricated for testing purposes. The overall system took two people 11 days to construct, and was mounted onto a temporary concrete wall in 1 day, as seen in Fig. 13. The design generated by the performative design environment was flawless. The ASF is currently undergoing tests to measure the electricity generation and adaptive control strategy.
The authors would like to acknowledge the HiLo project members for the design and construction of the ASF: Supermanoeuvre (Sydney Australia) and the Professorship of Architecture and Structures (BRG, ETH Zurich) for their work in designing the HiLo building. The authors would also like to thank other key contributors to the ASF Project: Bratislav Svetozarevic, Moritz Begle, and Anja Willmann. This research project is financially supported by the Swiss Innovation Agency Innosuisse and is part of the Climate Competence Center for Energy Research SCCER FEEB&D. In addition, this research has been financially supported by the Building Technologies Accelerator program of Climate-KIC. References [1] R. Loonen, M. Trčka, D. Cóstola, J. Hensen, Climate adaptive building shells: state-of-the-art and future challenges, Renew. Sust. Energ. Rev. 25 (2013) 483–493, http://dx.doi.org/10.1016/j.rser.2013.04.016. [2] Z. Nagy, B. Svetozarevic, P. Jayathissa, M. Begle, J. Hofer, G. Lydon, A. Willmann, A. Schlueter, The Adaptive Solar Facade: from concept to prototypes, Front. Archit. Res. 5 (2) (2016) 143–156, http://dx.doi.org/10.1016/j.foar.2016. 03.002. [3] P. Jayathissa, M. Luzzatto, J. Schmidli, J. Hofer, Z. Nagy, A. Schlueter, Optimising building net energy demand with dynamic BIPV shading, Appl. Energy 202 (2017) 726–735, http://dx.doi.org/10.1016/j.apenergy.2017.05.083. [4] A. Schlueter, F. Thesseling, Building information model based energy/exergy performance assessment in early design stages, Autom. Constr. 18 (2) (2009) 153–163, http://dx.doi.org/10.1016/j.autcon.2008.07.003. [5] R. Volk, J. Stengel, F. Schultmann, Building Information Modeling (BIM) for existing buildings - literature review and future needs, Autom. Constr. 38 (2014) 109–127, http://dx.doi.org/10.1016/j.autcon.2013.10.023. [6] R. Oxman, Performance-based design: current practices and research issues, Int. J. Archit. Comput. 6 (1) (2008) 1–17, http://dx.doi.org/10.1260/
4. Discussion and conclusion This paper presents a practical PDE for the design and fabrication of kinetic architectural elements. The PDE is capable of combining the multiple fields of structural engineering, energy engineering, control engineering, industrial design and architecture into one integrated environment, allowing the designers to handle the complexities of a multidisciplinary project. The major advantage of this environment is the considerable decrease in time between design iterations. Traditionally each of the respective stakeholders in the project would work on their individual design, and then exchange information in a meeting. With the PDE, the meeting is transformed from an information exchange session, to a 346
Automation in Construction 93 (2018) 339–347
P. Jayathissa et al.
147807708784640090. [7] M. Turrin, P. Von Buelow, A. Kilian, R. Stouffs, Performative skins for passive climatic comfort: a parametric design process, Autom. Constr. 22 (2012) 36–50, http://dx.doi.org/10.1016/j.autcon.2011.08.001. [8] M. Rippmann, L. Lachauer, P. Block, Interactive vault design, Int. J. Space Struct. 27 (4) (2012) 219–230, http://dx.doi.org/10.1260/0266-3511.27.4.219. [9] D. Holzer, Design exploration supported by digital tool ecologies, Autom. Constr. 72 (2016) 3–8, http://dx.doi.org/10.1016/j.autcon.2016.07.003. [10] D.Y. Kim, S.A. Kim, An exploratory model on the usability of a prototyping process for designing of Smart Building Envelopes, Automat. Constr. 81 (2017) 389–400, http://dx.doi.org/10.1016/j.autcon.2017.03.012. [11] H. Díaz, L.F. Alarcón, C. Mourgues, S. García, Multidisciplinary design optimization through process integration in the AEC industry: strategies and challenges, Autom. Constr. 73 (2017) 102–119, http://dx.doi.org/10.1016/j.autcon.2016.09. 007. [12] K. Negendahl, Building performance simulation in the early design stage: an introduction to integrated dynamic models, Autom. Constr. 54 (2015) 39–53, http:// dx.doi.org/10.1016/j.autcon.2015.03.002. [13] P. Block, A. Schlueter, D. Veenendaal, J. Bakker, M. Begle, J. Hofer, P. Jayathissa, G. Lydon, I. Maxwell, T. Mendez Echenagucia, Z. Nagy, D. Pigram, B. Svetozarevic, R. Torsing, J. Verbeek, A. Willmann, NEST HiLo: Research & innovation unit for lightweight construction and building systems integration, J. Build. Eng. 12 (2017) 332–341, http://dx.doi.org/10.1016/j.jobe.2017.06.013. [14] G. Lydon, J. Hofer, B. Svetozarevic, Z. Nagy, A. Schlueter, Coupling energy systems with lightweight structures for a net plus energy building, Appl. Energy 189 (2017) 310–326, http://dx.doi.org/10.1016/j.apenergy.2016.11.110. [15] N. Jakica, A. Zanelli, optical and energy yield co-simulation of new generation BIPV envelope in early design phase using custom ray tracing algorithm in python,
[16] [17] [18] [19] [20]
[21]
[22] [23]
[24]
[25] [26]
347
energy forum, Economic forum, 2014, pp. 1031–1038 doi:hdl.handle.net/11311/ 966160. Rhinoceros V5 (2015) [last accessed 2017-04-11]. https://www.rhino3d.com/. Grasshopper - Algorithmic Modelling for Rhino (2015) [last accessed 2017-04-11]. http://www.grasshopper3d.com/. Python [last accessed 2017-04-11]. https://www.python.org/. M.S. Roudsari, M. Pak, A. Smith, Ladybug, (2013) http://www.grasshopper3d.com/ group/ladybug. J. Hofer, A. Groenewolt, P. Jayathissa, Z. Nagy, A. Schlueter, Parametric analysis and systems design of dynamic photovoltaic shading modules, Energy Sci. Eng. 4 (2) (2016) 134–152, http://dx.doi.org/10.1002/ese3.115. EN ISO 13790: Energy Performance of Buildings: Calculation of Energy Use for Space Heating and Cooling. (2008) [last accessed 2017-04-11] https://www.iso. org/standard/41974.html. S.V. Szokolay, Environmental Science Handbook for Architects and Builders, Construction Press, Lancaster, 1980 ISBN 0860958132. P. Jayathissa, M. Luzzatto, J. Schmidli, ASF Simulation Framework, (2017) https:// github.com/architecture-building-systems/ASF_Simulation [last accessed 201704-11]. P. Jayathissa, RC Building Simulator, (2017) https://github.com/architecturebuilding-systems/RC_BuildingSimulatorhttp://dx.doi.org/10.5281/zenodo. 1172150. karamba 3d [last accessed 2017-04-11] http://www.karamba3d.com/. B. Svetozarevic, Z. Nagy, J. Hofer, D. Jacob, M. Begle, E. Chatzi, A. Schlueter, SoRoTrack: A two-axis soft robotic platform for solar tracking and building-integrated photovoltaic applications, Robotics and Automation, 2016 Institute of Electrical and Electronics Engineers International Conference, IEEE, 2016, pp. 4945–4950, , http://dx.doi.org/10.1109/ICRA.2016.7487700.
Contents lists available at ScienceDirect
Automation in Construction journal homepage: www.elsevier.com/locate/autcon
Performative design environment for kinetic photovoltaic architecture P. Jayathissa a b
a,*,1
, S. Caranovic
a,1
a
b
, J. Hofer , Z. Nagy , A. Schlueter
T
a
Architecture and Building Systems, Institute of Technology in Architecture, Department of Architecture, ETH Zurich, Switzerland Intelligent Environments Laboratory, Department of Civil, Architectural and Environmental Engineering, The University of Texas at Austin, USA
A R T I C LE I N FO
A B S T R A C T
Keywords: Kinetic architecture Multi-functional envelope Parametric design Adaptive shading Performative design
The design of complex architectural components such as kinetic architectural elements poses a challenge due to the multiple technological branches involved. This paper presents a performative design environment that combines the branches of structural and energy engineering, control, industrial design, and architecture. The methodology is applied in the context of the Adaptive Solar Facade, a kinetic photovoltaic shading system for the HiLo building in Duebendorf, Switzerland. The authors describe how the environment enables the user to design the form of the facade, get feedback on its structural strength, analyse the energetic performance of the interior space, conduct a daylighting analysis, render images, and produce fabrication plans for a rapid design process. With the parametric design environment, project meetings transform from information exchanges to design sessions where all stakeholders can collaboratively influence the design and see immediate results. What would normally take a month, was condensed to just a few hours. Ultimately, this paper extends the field of performative design by presenting a practical example where a system as complex as a kinetic photovoltaic envelope can be designed, prototyped, and fabricated by a small team of four designers.
1. Introduction The interest for applying kinetic envelopes in architecture has increased in recent years. On the one hand, there is the availability of digital design processes, which enables kinetic concepts to come to life. On the other hand, there is the desire to have active building elements to regulate energy flows and to create a more comfortable built environment [1]. When analysing the various energy flows, it is solar radiation that is of particular interest. As seen in Fig. 1, the mediation of solar radiation has the potential to reduce heating and cooling demands, while simultaneously distributing daylight according to the occupants' desires [2]. Furthermore, utilising thin film photovoltaic panels as kinetic elements enables the facade to also act as an electricity generator. Such technologies can, in some cases, offset the entire energy consumption of the rooms behind the envelope [3]. The application of kinetic architectural envelopes has so far been centred around iconic examples. These include the Al Bahar Towers in Dubai, the Arab World Institute in Paris, and the ThyssenKrupp Headquarters in Essen. Bringing these technologies to the mainstream, however, can be challenging.
One major challenge in the design of Kinetic Architectural Elements is the involvement of multiple technological branches. Among these branches one can count structural and energy engineering, control, industrial design, and architecture. Each of these branches has strong interdependencies to each other. For example, during the detail design process, a small variation in the control system, such as the range of movement, can affect the energetic performance and the architectural image. This may result in a redesign, that then needs to undergo structural evaluations and fit within the budget. These iteration cycles become a time intensive process that, in some cases, can take months to solve. In order to unite these branches efficiently, new design methods and environments have to be developed. The state of the art in interdisciplinary building design lies in building information modelling (BIM) [4,5]. The utilisation of BIM to perform fast energy and structural assessments can coordinate the design from its early stages. However, design using BIM is often based on high levels of standardisation, which makes it complicated when designing custom innovative components. The development of kinetic architectural envelopes requires a flexible infrastructure, which allows for fast design, prototyping and production, while maintaining the ability to be customised for each system individually.
Abbreviations: ASF, Adaptive Solar Facade; BIM, Building information modelling; CAD, Computer aided design; PDE, Performative design environment; PV, Photovoltaic; FEA, Finite element analysis * Corresponding author. E-mail addresses: [email protected] (P. Jayathissa), [email protected] (S. Caranovic), [email protected] (J. Hofer), [email protected] (Z. Nagy), [email protected] (A. Schlueter). 1 Equally contributing first authors. https://doi.org/10.1016/j.autcon.2018.05.013 Received 13 October 2017; Received in revised form 16 April 2018; Accepted 9 May 2018 0926-5805/ © 2018 Elsevier B.V. All rights reserved.
Automation in Construction 93 (2018) 339–347
P. Jayathissa et al.
shape the form of a facade, provide feedback on its structural strength, analyse the energetic performance and day-lighting conditions of the building space, render images, and produce fabrication drawings for a rapid iterative design process. This is accomplished within the Rhino/ Grasshopper environment with python as a scripting language. The methodology is applied in the context of the Adaptive Solar Facade (ASF), a kinetic photovoltaic facade constructed for the research and innovation unit known as the HiLo [13]. HiLo is a two bedroom apartment with a portfolio of energy saving technologies that create a net zero energy building [14]. Both bedrooms of the HiLo will be equipped with an ASF as seen in Fig. 2. The adaptive control of solar radiation into the bedrooms, coupled with on site electricity generation will contribute to this overall net zero energy strategy. The entire design process from conceptual design, to prototyping, and final fabrication will be presented. The remainder of the paper is organised as follows: the next section details the parametric design environment and the simulations that drive the final design. Section 3 describes the outcomes of this design process including the details of the kinetic facade, and in Section 4 the limitations of the environment are discussed and the paper is concluded.
importance max PV
ASF min
daylighting
views
Cooling
ASF Fig. 1. The facade acting as a mediator between the interior and exterior environment, while fulfilling various functions [2,15].
Performative design environments can overcome the limitations of BIM by taking computer aided design (CAD) in a reversed manner, where it is the simulations that drive the design [6]. Here, the concept of form making, is replaced with form finding. An example of performative design has been described by Turrin et al. where passive solar strategies were explored to improve the thermal comfort and daylight quality under a tessellated roof [7]. This can also stretch to structural performative design in tools such as RhinoVault where the final form is determined through iterative structural simulations [8]. Holzer on the other hand used performative design for direct structural feedback using first and second order structural simulations [9]. These tools can also be utilised on the component level, and have been previously analysed in a design studio for smart building envelopes [10]. However, integrating multiple tools in a single automated environment has still proven to be difficult due to some tools lacking parametrisation capabilities, low openness of the tools' interfaces, and low flexibility [11,12]. Furthermore, this lack of interoperability results in long iteration cycles, making it difficult to evaluate the necessary trade-offs. In this paper, the authors build on this existing knowledge to produce a state of the art performative design environment (PDE) that can
2. Methodology The parametric design environment, shown schematically in Fig. 3 details the multiple processes that were used in parallel to handle the different technological branches. From this, four key outputs are generated: the structural performance, energetic performance, manufacturing plans, and visual renders. Inputs to the PDE are defined by the parameters that have the greatest influence on the design. In this case, they are the overall frame dimension and profile, the photovoltaic panel dimension, spacing and layout, and the range of motion. These inputs are numerically fed into the design environment and generate instantaneous results of the structural performance, energetic performance, visual renders and manufacturing plans. By doing so, the multiple trade-offs between the technological branches, as explained in Section 1, can be simultaneously observed. The electricity generation, building energy demand, utilisation factor of yield strength,
Fig. 2. Render of the HiLo module to be constructed in Dubendorf, Switzerland. Two adaptive solar facades have been planned on the south and west facing facades. 340
Automation in Construction 93 (2018) 339–347
P. Jayathissa et al.
Fig. 3. The performative design environment is able to link three analysis methodologies, resulting in rapid design iterations. Here the panel spacing input parameter of the environment is varied. Once a design has been selected, production plans can be automatically generated leading to rapid physical implementation.
structural and energetic analysis in more detail.
dimensions, collision detection, aesthetics, and manufacturing costs are of particular interest. This ultimately allows for quick feedback loops with the major stakeholders (architects, structural engineering, energy engineers, and production team) involved in the project. The environment combines the geometric modelling software Rhinoceros 3D [16], its parametric plug-in Grasshopper [17], and python [18] as a scripting language. The relatively unspecialised nature of Rhino is complemented by a large number of specialised add-ons for Grasshopper. Furthermore, custom Grasshopper modules can be scripted using python. The following subsections will explain the
2.1. Energy evaluation The purpose of the ASF is to maximise electricity production on the photovoltaic (PV) panels and minimise the energy consumption of the building behind the facade. In order to best achieve this, the layout of the PV panels, and the electrical interconnection of PV panels must be carefully designed. The evaluation of the energetic performance of the ASF can be 341
Automation in Construction 93 (2018) 339–347
P. Jayathissa et al.
panels enables more control over the solar insolation. Tight configurations tend to perform better in hot climates, whereas sparse configurations perform better in colder climates. 4. Daylighting model: A linear daylighting model based on the total flux methodology is used to determine the luminosity in the room [22]. When the luminosity falls below a threshold value of 300lx, artificial lighting is turned on. Sparse configurations result in more daylight distribution, which reduces the need for artificial lighting in the morning and evening hours. 5. Optimisation: The simulation is conducted for all possible panel angle combinations for every hour of the year. By summing all the time steps, the annual energetic performance of the ASF can be evaluated.
Fig. 4. A simulation result showing the radiation on the solar panels and the window element on 16 June 2013 between 12:00 and 13:00 in Zurich.
This analysis finds the optimum balance between PV generation and daylight control to minimise heating, cooling and lighting demands where the overall objective is the minimisation of net energy. The source code for this methodology, with installation instructions can be downloaded from github [23,24]. During the design stage, this analysis is conducted for a typical day in summer, and in winter. Once a design has been selected, an annual study with hourly time steps is conducted to achieve a high resolution result.
found in [3], and will be briefly reviewed here for completion. This part of the PDE consists of five stages: 1. Solar radiation model: The radiation on the PV panels and window behind the ASF is calculated using the Grasshopper - LadyBug plugin. [19]. An example of the simulation result is shown in Fig. 4. A tighter layout of panels allows for more PV material per facade area, however it also results in more module self-shading, which reduces the overall electricity production. 2. PV electricity production: The radiation result on the PV panels is coupled to an electrical circuit simulation of monolithically interconnected, thin-film CIGS PV modules. This model takes into account the effects of module self-shading and temperature dependence [20]. 3. Building energy model: The radiation calculated on the window surface is fed to a 5R1C single zone resistance-capacitance building model based on the ISO 13790 standard [21]. This calculates the heating or cooling demand of the building. A tight layout of PV
2.2. Structural form finding and finite element analysis The panel spacing, as discussed in Section 2.1 can also influence the structural performance of the system, which in turn, influences the architectural image. It is therefore important to run the structural analysis in parallel to find the optimum solution. This subsection will first introduce the structural concept, and then detail how this concept was developed with the PDE. The proposed ASF is based on a vaulted cross-hatched network of stainless steel pipes as seen in Fig. 5. The junctions where the pipes
Fig. 5. The Adaptive Solar Facade existing in varying states. 342
Automation in Construction 93 (2018) 339–347
P. Jayathissa et al.
Fig. 6. An example output from the Karamba structural simulation. The black arrows detail the loading direction, and the colour details the utilisation factor in relation to the von Mises stress. The table on the right details the input parameters to the simulation.
Fig. 7. Effect of annual PV production with respect to module spacing per square meter of facade area [20].
cross serve as the mounting points for each dynamic photovoltaic module. All utility lines are routed within the pipe network. A steel frame supports the pipe network to create a stand-alone pre-fabricated component, which can be mounted directly to the building envelope. The vaulted shape further strengthens the structure against wind loads, thus allowing for thinner pipe diameters, which increases transparency. The form of the vaulted structure is determined through RhinoVault, a form finding plug-in for Grasshopper [8]. The method calculates the optimum shape of the pipe network based on the loading points and inputs to the design environment. Once the form is determined, a structural second order finite element analysis is conducted using the Karamba3d plug-in [25] to dimension the structural elements. Further manual adjustments to the mesh can be conducted to improve the architectural image. Each manual adjustment is directly computed by a second order finite element analysis (FEA), providing real time feedback about the stability of the adjusted structure. Fig. 6 shows an example of this output with a list of simulation parameters. A load of 420N was applied on each panel node, which is equivalent to a category one hurricane on the Saffir-Simpson scale. The results are visualised in the form of coloured meshes which detail the utilisation factor in relation to the von Mises stress. A scaled down 2.2 m × 1.5 m prototype was constructed to validate the structural model and the design of the connection details.
Fig. 8. Utilisation factor, represented as a percentage of the maximum yield strength. Failure occurs with utilisation factors greater than 100%. The utilisation factor decreases with increasing frame size cross section, and with greater vaulted depth.
2.3. Integration of classical design methods A purely parametric approach is very useful in investigating design possibilities, but once the concept needs to be put in practice, the parametric design environment needs to be complemented with 343
Automation in Construction 93 (2018) 339–347
P. Jayathissa et al.
Fig. 9. Prototype constructed for model validation. a, b details the experimental set up. c, d detail the snap-through buckling of the structure. Table 1 Comparison of the physical and simulated model for the scaled down prototype.
Max point loading Max deflection
Physical model
Simulated model
Deviation
250 N 14.5 mm
220 N 10.3 mm
−12% −29%
Fig. 11. Exploded view of the final ASF module.
physical prototypes, testing and classical design methods like simple 3D modelling and electronics design. As there is a trade-off between the flexibility of a parametrised environment and the time required for its development, it is important to wisely choose what to parametrise and what to design with classical methods. Parametrising the entire design process of complex systems such as an ASF can result in instabilities and lead to an increase in design time. For this reason, it is important to include certain constants amongst the parametric variables. Constants include the design of the electronics, connection details, and the actuator design [26]. Fig. 10. Constraints of panel motion relative to the structure.
3. Results This section details some of the outcomes of the aforementioned 344
Automation in Construction 93 (2018) 339–347
P. Jayathissa et al.
3.2. Optimum supporting frame dimension With a frame size of 2.5×4.02 m, a PV panel spacing of 510 mm and a structural depth of the vault of 150 mm, the shape of the rod-net vault, the profiles of the pipes, and the size of the supporting frame were calculated. With a wind load of 0.92 KN /m2, and a safety factor of 1.8, the finite element analysis concluded that the stainless steel pipe and the rectangular steel frame have a minimum dimension of 16×2 mm and 180× 120×6 mm respectively. The relatively large frame is the result of a snap-through buckling failure criterion and relatively weak support conditions at the corners. As the vault is loaded, the large lateral forces on the frame result in deflections, which in turn reduce the depth of the vaulted structure. The reduced depth of the vault further increases the lateral loading, thus leading to a global snap-through buckling failure. The frame must therefore be large enough to withstand the maximum load criterion with minimal deflection. This solution is characteristic to the conditions present on the HiLo building, where the ASF will be mounted with an offset of 80 cm from the building's support structure. In principle, the edge points of the pipe structure could be mounted directly to the support structure of the building, thus reducing the need for a frame. Fig. 8a depicts how the utilisation of the pipe elements decreases with increasing frame strength, up to the point where all edge points are supported individually. The 150 mm depth of the vaulted structure is sufficient to handle the wind loads with relatively thin pipes and minimal variation in the panel orientation. Ultimately, the choice of the structural depth is a trade-off between stability, self-shading, and aesthetics. A deeper structure offers more load bearing capacity requiring thinner pipes, however the self-shading of the panels increase. On the other hand, a shallow structure is less resilient to wind loads, leading to thicker pipes. The effect of the depth on the structural stability is depicted in Fig. 8b. The figure shows how the maximum utilisation of the pipes decreases with increasing structural depth, with a very steep decrease in the range of 12–15 cm. The construction details contribute significantly to the overall stability of the structure and to the match between the real behaviour and the FEA simulation. For this reason, the authors designed the pipe and frame connections to provide a completely stiff connection under the assumed wind loads. To form the vaulted shape, each pipe element is bent close to its edges and kept straight in between those points to improve their stability to localised buckling.
Fig. 12. Automatically generated drawings for pipe bending. Here four of the 92 pipe drawings are shown.
design environment in relation to the case study at the HiLo building. In particular the optimal PV panel layout, the structure, details of the module design, and fabrication will be evaluated. A typical iteration of the PDE would take 1 min. However a high resolution hourly analysis of the energetic performance with all solar angles can take up to 6 h. During a typical meeting, the simplified model would be run with approximately 20 iterations. Afterwards, a high resolution model is run for validation.
3.1. Optimum PV panel layout The design of the PV panel layout is an example where all four stakeholders had inputs to the design. From an energetic perspective, a dense layout would have a larger overall PV surface area for electricity generation. However if the panels are too close together, there would be high module self shading which would lower the overall performance of the panels, as seen in Fig. 7 [20]. Furthermore, there would be less natural lighting in the room, resulting in an increase of the building's overall energy consumption. From an architectural perspective, a sparse PV layout is preferred as it increases the transparency of the facade from the inside. This also lowers the overall cost, as less PV panels are required. Structurally, however, a sparse PV layout results in longer pipe elements between the junctions, which in some configurations, decreases the overall strength due to an increased buckling length. Ultimately, a PV panel spacing of 510 mm with a module size of 420 mm was chosen as a trade-off between the above-mentioned requirements.
3.3. Structural validation Due to the complexity of the joints and double curved form, it was
Fig. 13. Mounting of a prefabricated ASF to the test site. 345
Automation in Construction 93 (2018) 339–347
P. Jayathissa et al.
design session where all stakeholders can collaboratively influence the design and immediately see the necessary results. With traditional methods, five design iterations would normally take a month, whereas with the PDE, this can be condensed to a few hours. The design environment also develops with the project allowing for new parametric inputs, or new outputs to be created. The management of such a parametric design environment can be easily done with collaborative software management tools such as git, a distributed version control system. One disadvantage is the overhead required to manage a PDE. Like a BIM manager, a PDE manager is required and must have a careful overview of the software. As the software ultimately determines the final form of the design, any errors in the software can be detrimental to the final design. It was therefore necessary for all stakeholders to conduct a final independent analysis prior to the submission of the final design. One limiting factor in the design of the PDE is the computational time. The full annual energetic analysis, for example, may take 6 h to solve. Simplifications were therefore made to accelerate this process during the design stage, and the full complex evaluation was conducted afterwards to validate the simplified model. It is also important to determine what aspects of the design should be contained within the PDE and what should be designed with classical methods. Essentially, all aspects where there could be conflicts between the stakeholders were included in the PDE, whereas many of the design details, such as the mounting brackets to the building, can still be designed independently and imported into the PDE as a static object. Over time, some static objects, such as the cantilevered bracket of the PV module became parametrised. Ultimately, this paper presents a further step in the field of performative design by showing how a system as complex as a kinetic photovoltaic envelope, can be developed, prototyped and finally fabricated by a small team of four designers. The methodology can be utilised for any building component where multiple technological branches are required.
important to validate the structural analysis. A scaled down, 2.2 m× 1.5 m, prototype designed through the PDE was constructed for testing purposes. Weights were added to each of the seven junction nodes in 5 kg increments until the structure collapsed as seen in Fig. 9. Each junction node was also fitted with a deflection gauge. The results of this analysis are summarised in Table 1. The simulated model underestimates the overall strength by 12% which is sufficient to validate our model. The deflection results, however, have a larger diversion. This diversion can be attributed to a weak construction joint at one of the edge nodes. The average deflection of the structure is 10.5 mm which is close to the simulated model. Besides confirming the accuracy of the simulation, this test also confirmed the predicted snapthough buckling failure. Additionally, as seen in Fig. 9d, the junction nodes, and frame connections were capable of resisting local torsion and maintained their position. The failure was localised at the pipes as predicted in Section 3.2. 3.4. Module design The kinetic PV module was then parametrically designed to fit within the requirements of the optimised layout and structure. The PV panels must be able to fully open, and fully close without making contact with the rod-net structure or other PV modules as shown in Fig. 10. The PV panel is actuated using a soft pneumatic actuator [26]. The actuator is made from neoprene rubber and contains three air chambers. By pumping one or more of these chambers with compressed air, the actuator will deform, thus moving the panel with a 90° range in two degrees of freedom. The cantilevered bracket connects this actuator to the rod-net structure and holds it at a distance that prevents collision with other PV panels or the structure. A decentralised control box, located behind the junction element, contains three pneumatic valves and an electronic board addresses the valves over a common data bus. This controls the flow of air to the chambers inside the actuator. An exploded view of the module is shown in Fig. 11. 3.5. Fabrication
Acknowledgements Once a design has been configured using the PDE, manufacturing plans can be automatically generated. This enables the plans to be immediately sent to manufacturers, without a significant time investment. This is especially important for components such as the steel pipes where each pipe has a unique length and bend angle. As an example, Fig. 12 details four of the 92 pipe bending plans. Besides the plan, the PDE also generates a set of meta data that can be directly fed into a computer controlled pipe bending machine, thus simplifying the transition from design to production. An ASF for the HiLo building was fabricated for testing purposes. The overall system took two people 11 days to construct, and was mounted onto a temporary concrete wall in 1 day, as seen in Fig. 13. The design generated by the performative design environment was flawless. The ASF is currently undergoing tests to measure the electricity generation and adaptive control strategy.
The authors would like to acknowledge the HiLo project members for the design and construction of the ASF: Supermanoeuvre (Sydney Australia) and the Professorship of Architecture and Structures (BRG, ETH Zurich) for their work in designing the HiLo building. The authors would also like to thank other key contributors to the ASF Project: Bratislav Svetozarevic, Moritz Begle, and Anja Willmann. This research project is financially supported by the Swiss Innovation Agency Innosuisse and is part of the Climate Competence Center for Energy Research SCCER FEEB&D. In addition, this research has been financially supported by the Building Technologies Accelerator program of Climate-KIC. References [1] R. Loonen, M. Trčka, D. Cóstola, J. Hensen, Climate adaptive building shells: state-of-the-art and future challenges, Renew. Sust. Energ. Rev. 25 (2013) 483–493, http://dx.doi.org/10.1016/j.rser.2013.04.016. [2] Z. Nagy, B. Svetozarevic, P. Jayathissa, M. Begle, J. Hofer, G. Lydon, A. Willmann, A. Schlueter, The Adaptive Solar Facade: from concept to prototypes, Front. Archit. Res. 5 (2) (2016) 143–156, http://dx.doi.org/10.1016/j.foar.2016. 03.002. [3] P. Jayathissa, M. Luzzatto, J. Schmidli, J. Hofer, Z. Nagy, A. Schlueter, Optimising building net energy demand with dynamic BIPV shading, Appl. Energy 202 (2017) 726–735, http://dx.doi.org/10.1016/j.apenergy.2017.05.083. [4] A. Schlueter, F. Thesseling, Building information model based energy/exergy performance assessment in early design stages, Autom. Constr. 18 (2) (2009) 153–163, http://dx.doi.org/10.1016/j.autcon.2008.07.003. [5] R. Volk, J. Stengel, F. Schultmann, Building Information Modeling (BIM) for existing buildings - literature review and future needs, Autom. Constr. 38 (2014) 109–127, http://dx.doi.org/10.1016/j.autcon.2013.10.023. [6] R. Oxman, Performance-based design: current practices and research issues, Int. J. Archit. Comput. 6 (1) (2008) 1–17, http://dx.doi.org/10.1260/
4. Discussion and conclusion This paper presents a practical PDE for the design and fabrication of kinetic architectural elements. The PDE is capable of combining the multiple fields of structural engineering, energy engineering, control engineering, industrial design and architecture into one integrated environment, allowing the designers to handle the complexities of a multidisciplinary project. The major advantage of this environment is the considerable decrease in time between design iterations. Traditionally each of the respective stakeholders in the project would work on their individual design, and then exchange information in a meeting. With the PDE, the meeting is transformed from an information exchange session, to a 346
Automation in Construction 93 (2018) 339–347
P. Jayathissa et al.
147807708784640090. [7] M. Turrin, P. Von Buelow, A. Kilian, R. Stouffs, Performative skins for passive climatic comfort: a parametric design process, Autom. Constr. 22 (2012) 36–50, http://dx.doi.org/10.1016/j.autcon.2011.08.001. [8] M. Rippmann, L. Lachauer, P. Block, Interactive vault design, Int. J. Space Struct. 27 (4) (2012) 219–230, http://dx.doi.org/10.1260/0266-3511.27.4.219. [9] D. Holzer, Design exploration supported by digital tool ecologies, Autom. Constr. 72 (2016) 3–8, http://dx.doi.org/10.1016/j.autcon.2016.07.003. [10] D.Y. Kim, S.A. Kim, An exploratory model on the usability of a prototyping process for designing of Smart Building Envelopes, Automat. Constr. 81 (2017) 389–400, http://dx.doi.org/10.1016/j.autcon.2017.03.012. [11] H. Díaz, L.F. Alarcón, C. Mourgues, S. García, Multidisciplinary design optimization through process integration in the AEC industry: strategies and challenges, Autom. Constr. 73 (2017) 102–119, http://dx.doi.org/10.1016/j.autcon.2016.09. 007. [12] K. Negendahl, Building performance simulation in the early design stage: an introduction to integrated dynamic models, Autom. Constr. 54 (2015) 39–53, http:// dx.doi.org/10.1016/j.autcon.2015.03.002. [13] P. Block, A. Schlueter, D. Veenendaal, J. Bakker, M. Begle, J. Hofer, P. Jayathissa, G. Lydon, I. Maxwell, T. Mendez Echenagucia, Z. Nagy, D. Pigram, B. Svetozarevic, R. Torsing, J. Verbeek, A. Willmann, NEST HiLo: Research & innovation unit for lightweight construction and building systems integration, J. Build. Eng. 12 (2017) 332–341, http://dx.doi.org/10.1016/j.jobe.2017.06.013. [14] G. Lydon, J. Hofer, B. Svetozarevic, Z. Nagy, A. Schlueter, Coupling energy systems with lightweight structures for a net plus energy building, Appl. Energy 189 (2017) 310–326, http://dx.doi.org/10.1016/j.apenergy.2016.11.110. [15] N. Jakica, A. Zanelli, optical and energy yield co-simulation of new generation BIPV envelope in early design phase using custom ray tracing algorithm in python,
[16] [17] [18] [19] [20]
[21]
[22] [23]
[24]
[25] [26]
347
energy forum, Economic forum, 2014, pp. 1031–1038 doi:hdl.handle.net/11311/ 966160. Rhinoceros V5 (2015) [last accessed 2017-04-11]. https://www.rhino3d.com/. Grasshopper - Algorithmic Modelling for Rhino (2015) [last accessed 2017-04-11]. http://www.grasshopper3d.com/. Python [last accessed 2017-04-11]. https://www.python.org/. M.S. Roudsari, M. Pak, A. Smith, Ladybug, (2013) http://www.grasshopper3d.com/ group/ladybug. J. Hofer, A. Groenewolt, P. Jayathissa, Z. Nagy, A. Schlueter, Parametric analysis and systems design of dynamic photovoltaic shading modules, Energy Sci. Eng. 4 (2) (2016) 134–152, http://dx.doi.org/10.1002/ese3.115. EN ISO 13790: Energy Performance of Buildings: Calculation of Energy Use for Space Heating and Cooling. (2008) [last accessed 2017-04-11] https://www.iso. org/standard/41974.html. S.V. Szokolay, Environmental Science Handbook for Architects and Builders, Construction Press, Lancaster, 1980 ISBN 0860958132. P. Jayathissa, M. Luzzatto, J. Schmidli, ASF Simulation Framework, (2017) https:// github.com/architecture-building-systems/ASF_Simulation [last accessed 201704-11]. P. Jayathissa, RC Building Simulator, (2017) https://github.com/architecturebuilding-systems/RC_BuildingSimulatorhttp://dx.doi.org/10.5281/zenodo. 1172150. karamba 3d [last accessed 2017-04-11] http://www.karamba3d.com/. B. Svetozarevic, Z. Nagy, J. Hofer, D. Jacob, M. Begle, E. Chatzi, A. Schlueter, SoRoTrack: A two-axis soft robotic platform for solar tracking and building-integrated photovoltaic applications, Robotics and Automation, 2016 Institute of Electrical and Electronics Engineers International Conference, IEEE, 2016, pp. 4945–4950, , http://dx.doi.org/10.1109/ICRA.2016.7487700.